Team Deadlock

Copy of the repository for the team Deadlock for Newsmatics Hackathon 2025

• Team captain: Erik Čapkovič
• Authors: Erik Čapkovič, Lukáš Denkócy, Rastistlav Kollár, Lukáš Medovič, Peter Štefunko

Prerequisites

Before you begin, ensure you have the following installed on your machine:

• Python 3.13+
• Poetry
• Docker
• Docker Compose
• PostgreSQL

Installation

1. Clone the repository:

   git clone https://github.com/peter-stefunko/team-deadlock
   cd team-deadlock

2. Install the dependencies using Poetry:

   cd backend
   poetry install

3. Set up the environment variables:

   Create a .env file in the root of the project and add the necessary environment variables. Refer to .env.template for the required variables.

Python venv

1. To develop in terminal, run the following command in the backend directory:

   source $(poetry env info --path)/bin/activate

2. If you are using an IDE for development like PyCharm or VSCode, refer to their own guide for setting up a Poetry environment.

Database Migration

1. To create a new migration, use:

   alembic revision

Running the Project

1.Firstly, you should build the whole project with: sh make build

2. You should now be able to start it up with:

   make start

3. The backend should now be running at http://localhost:8000. (with the swagger being at http://localhost:8000/api/docs)

   

4. You should be able to access the website at http://localhost.

5. For more make commands, run the following:

   make help

Usage

• Once you have it up and running, you can now use it to it's fullest potentiol. On the main screen of the website, there is a user search input where you can ask our system anything in you owns words, and it should find the most accurate article you are looking for.
• You can also use the random phrase generator on the left, if you feel lucky
• There are also filters on the left, and bellow the search bar, you can toggle addiotional categories, which are appended to your's search input for further enhancing similarity accuracy
• Every article is composed of the clickable title, trust score from 0 to 5, publisher, classification and also you can click up or downvote to change the article's trust score

Live Demo

We have a live demo of the project where you can explore the embedding-based search functionality in action.

• Demo URL: Live Demo
• Demo Swagger: Live Swagger
• Usage: Simply enter your query in the search bar, and our backend will retrieve the most relevant results using our optimized embedding-based search.
• Authentication: Some features may require authentication. If applicable, refer to the provided test credentials or register for access.
• Feedback: If you encounter any issues or have suggestions, please open an issue in this repository or contact us.

Check out the demo and experience our solution firsthand!